II

Edition 4 January 2000. Copyright 1995.All rights reserved.

No part of this publication may be reprintedwithout written permission from the Author.

III

Personage behind the research and the results in short.This book is the result of the relentless work and research

done by the retired boat yard owner Mr. Bengt Blomberg ofStockholm, Sweden and his team of laminate and coatingspecialists .

An offshore powerboat pilot (Offshore2 world speed recordholder 1969-93 together with his wife Heidi), active in themarine repair and boat building business for over fourdecades, Mr. Blomberg became familiar with the problem inthe early 1970's. During the following 20 years his yardperformed hundreds of osmosis repairs.

Over the years, he actively tested and recorded theperformance of just about every known "osmosis" repairmethod; from propane burners, to grinding, sandblasting,peeling and infrared heat.

Mr. Blomberg was the first to use high pressure (350 bars)wet-slurry blasting systems for gel coat and blister removal,but the long -term results continued to be disappointing.

The "osmosis" recurred to a greater or lesser degree andno method seemed to permanently cure the problem.

Initially, he accepted the explanations of the "osmosis"problem provided by resin and coatings manufacturers,surveyors and other "experts" and followed theirrecommended repair procedures.

Soon finding these methods and their associated theories tobe inadequate and illogical, Mr. Blomberg set out to find thetrue reasons for and the chemical facts behind the problem.

IV

A closer survey of all the many circulating different andconfusing explanations of the problem revealed more of anindustry that was anxious not to frighten presumptive boatbuyers, than presenting the truth of it.

1953-64 Mr Blomberg was director of a major automotivechemical factory where he among else developed the firstwater-soluble detergent for removing of silicone contaminatesfrom windscreens and delicate enamels. The formula is stillwidely used when cars are re-painted. Also a co-founder of theSwedish-German-Polish-US rust prevention research firm SafeCoat, Mr Blomberg had sufficient chemical knowledge andtrade connections to be able to scrutinise the confusinginformation available about "osmosis" and polyester relatedresearch.

A number of hulls with different stages of osmosisdamages were bought and cut into pieces for laboratory testsand strength comparisons.

Within a couple of years the team could present the exactreasons for and the chemistry behind the problem:

A polyester laminate needs post curing at 80-90C in orderto cure completely. Of economical reasons this is, with fewexceptions, not performed at boat construction.

A not post cured laminate contains 10-15% uncured and nottotally cured polyester resin. As described in the followingpublication, such uncured resin enclosures can be transformedinto its original acid-alkali-glycol substances by adding of a fewwater molecules.

This is called hydrolysis, but has nothing to do with excesswater entrance or faulty construction, as commercial sourceshave brainwashed us to believe for many years!

VAll laminates, perfect or with faults, with or without specialwatershield coats, will absorb enough water to feed thehydrolysis. It is just a property of the material, which the mightyindustry has used all means to hide from the consumers.

At this stage the research team became confused: What wehad found was related to practically all produced GRP hulls.Why then were not all of them affected by the hydrolysis?

Could we be on the wrong way?

Another two years of numerous tests learned us a lot moreof the so called osmosis process but not what we werelooking for: The trigger.

1992 Mr. Blomberg by sheer coincidence met a retireddirector of research at a major British resin manufacturer.

The researcher was familiar with the problem, havingstudied it in the early 1960's, as architectural FRP panelsmysteriously had developed "lumps" or "warts" - what we todaywould call "osmosis" blisters.

The studies at the time led to a better post curing technique,which remedied the problem. The project was shelved and thefindings were never made public.

The information gleaned from this study suggested a veryimportant theory : Normally the uncured enclosures can notbe affected by the water.

Occasionally specific "trigger" enclosures of styrenecovered by uncured polyester and pierced by a fibreglassstrand exist.

The strand allows water molecules or even moisture toenter and react with the styrene diluted inside of the uncuredpolyester skin and the hydrolyse starts!

VI

In order for a styrene globule to form around a strand, andcreate such a "trigger", certain specific criteria induced duringthe manufacturing process had to be fulfilled.

Some laminates have no such "triggers" and will thereforenot develop "osmosis", explaining the random nature of theproblem.

( Note: The explanation of the enclosures has been keptsimple for the normal yacht owner, but as a few ones withchemical background have pointed out, that styrene will neverform drops in the curing laminate but pass through as a gas.Therefore I must explain it a little better. What causes theenclosures are small drops of resin and other substancesdissolved in styrene which of different reasons do not cure.After being enclosed within the cured laminate, the not curedresin and eventually other substances fall out against thecured "wall" and leave styrene as a fluid in the center.)

Thorough testing of the "trigger" theory was performedusing remaining material from the earlier used test hulls.When test pieces from areas just over the waterline was splitup into separate laminate layers, a few pin-head sized darkspots were found. The microscope clearly revealed enclosureswith both the penetrating strand and the hydrolysed resin.

After considerable time, effort and expense, this extensiveresearch confirmed the validity of the "trigger" theory, revealedthe true cause of the problem and worst of all, found that theprocess causes not only cosmetic blisters but also verysevere damages deep in the laminate.

VII

Important research results in short:

1. Totally contradicting earlier theories, theosmosis is not caused by excessive waterentrance. There is absolutely no osmotic forces involved(the gelcoat is an absorbing material which cannever act as an osmotic membrane)!

It is the hydrolysis process, which creates thewater soluble products and the cavities, which thenallows for excess water to enter and cause the highmoisture readings.

2. It is the pthtallic acid, formed in the process, which causes the chain reaction and damages, notthe water! The acid forms as not water soluble crystalsand can not be washed out of the laminate asclaimed by some epoxy suppliers. Even after years of hardstand and weeklywashes, the acid will remain in the capillaries andcavities, dissolved in styrene and glycol. The hydrolysis will restart as soon as the hull islaunched, how well watershielded it might be. The laminate can become severe damagesbefore any warning blisters show up again.

VIII

3. Long time tests proved: Contradicting earliertheories most osmosis hulls, conventionally driedand epoxy coated without effective acid removal,become deep damages faster than similar hulls nottreated at all! Excess water in the laminate does no harm!Instead, due to temperature differences in thesurroundings, the water moves in and out and thuswashes out a lot of the harmful residues. Further a free moisture movement in thelaminate has a cooling effect which considerablyslows down the hydrolysis.

The hysteria about watershielding is createdonly by the heavy advertising about perfect epoxyor other expensive (= high margin) products whichshould keep the water out. In reality no such product will do better than toslow down the water entrance 3-7 times moreeffective than the gelcoat.

Therefore it can not stop the hydrolyse. It willonly cover the problem in a way, that you will not sethe warning blisters for a long time! Instead the coating decreases the abovementioned washing and cooling effects andaccelerates the damages.

IX

A watershield coating without a totalremoval of acid remnants, styrene and uncuredresin is a waste of your money to-day and willcost you a lot more in the future!

If a proper treatment with controlled residueremoval (just moisture meter control is of no value)can be performed, a watershielding is advisable. Not in order to stop hydrolyse, but to avoidweight increase for as long as possible afterlaunching.

When such a treatment is too expensive inrelation to the value of the boat, our experience is: Do not watershield, just sand eventually blistersopen once a year and apply antifouling only. This will retain de-laminations and the softeningof the hull for much longer than an epoxy coating!

4. Osmosis blisters is really not a cosmetic problem only! They are just the visible sign that thehydrolysis of the polyester has affected at least onelaminate layer. Already when the first blister shows, hydrolysealkali products have reduced the bond betweenfibre glass and polyester in the total under waterlaminate! This causes a 20-30% loss of structural strengthwhich no dry and shield treatment can restore.

X Without a proper treatment large parts of thelaminate layers will sooner or later have thepolyester hydrolysed with only soft wet fibreglassleft.

5. During the research hundreds of larger hulls withheavy woven roving mats in the laminate havebeen surveyed.

All were found to have numerous triggersinside each roving layer and more or less severehydrolysis de-laminations deep in the laminate.

However most of them showed no telltaleblisters at all!

All owners of hulls of 45 and up aresincerely advised to use a Tramex Skippermoisture meter on scale 2 and check for highreading areas every 2-3 years.

Drilling of a few 10 mm test holes in suchareas will reveal damages before the repaircosts run too high.

XI

6. In order to permanently stop the hydrolysis andrestore the bond between glass and polyester twotasks were found inevitable.

First: Heating of the laminate to 80-90C.

A. In order to post cure as much as possible of theuncured enclosures.

B. To restore the bond between the glass and thepolyester.

C. To get all pthtallic acid dissolved and fluid.D. To make the cured polyester temporarily soft

and enable removal of deep situated hydrolyseresidues and remaining uncured polyester.

Second: A technique to remove at least thepart of the acid which is dissolved in styrene.

The acid dissolved in the glycol is much lessdangerous as the glycol is imprisoned in theoriginal enclosure cavities. The molecules are too large to move along thecapillaries. Of course the acid/glycol/water mixture found ineventually de-lamination areas must be removed.

XII

So far the research has not revealed any other technique,than the so called hyper absorption or short HYAB, whichcan perform all tasks needed without damaging thelaminate.

This technique is based upon alternating, high volume,hot and cool fast air streams directed against the laminatesurface. It has been used during the research since 1988 withperfect results. When the team later took over and improved the patentsinvolved, the name HYAB DryTech was adopted for thework.

1993-94 test program Centres were established in theworst osmosis hit areas of the World: New Zealand, Hong Kong, Florida and Canary Islands. About 25 badly damaged hulls with one or more earlierunsuccessful treatments in their histories were repaired ateach Centre in order to prove the developed technique.

The research team concentrated on finding techniquesalso for restoring the loss of strength and softness of thelaminate caused by the hydrolysis. The resulting drilling and epoxy injection technique wasapplied already on some of the above mentioned test hullswith perfect results. Independent University tests 1998 proved, that thetechnique made the laminate even stronger and stiffer thanoriginal!

At the end of 1998 over 80 of the test hulls had beenlaunched for over 4 years without any recurrence signs andalso the rest of them after 2-3 years had no problems.

XIII

1999 the rights of bringing the developed techniquesinto the market was transformed to the Dutch companyHYAB International b.v. It is marketed as HYAB Osmocure and besides takingover the South Pacific were now six HYAB Centres arebusy, the Company has established new Centres inHolland, Germany, Mallorca, Malta and Greece.Interested parties can get info from their web site:www.hyab-osmocure.com

However the research proceeds in order to find still betterand less costly means for hydrolyse treatment.

May be anyone, after studying the following detailedexplanation of the problem, can get any ideas about whatcan be done? Please then contact Mr Blomberg by e-mail:blomberg@osmosisinfo.com

As an example I can now, July 2001, present the HotVacsystem developed by the English team Terry Davey andJohn Ashton. It is based on heated vacuum mats and afterthorough tests I can verify, that it functions fully as well asHYAB and without any environmental problems. Also itneeds much less manual labor.Visit their web site www.hotvac.com

This was a little push for the so far only repair facilitieswhich can provide a total solution for the problem.Hopefully more will follow.

RegardsThe Author.

XIV

XV

CONTENTS

CHAPTER 1The chemical processes involved and their influenceon laminate strength.

1. WHAT CAUSES "OSMOSIS"? ...........................................3

1.1. "Osmosis" is in reality not an adequate name.............3

1.2. Osmosis reason: Hydrolyse of uncured resin!.........4

1.3. Uncured resin in laminate can be hydrolysed..............7

1.4. Why does the "osmosis" not start in all FRP hulls?...7

1.5. Trigger found: fibre strand penetrating enclosure.......7

1.6. Construction faults influence process severity...........9

2.4. Woven roving creates more triggers...........................13

2.5. Some blisters are harmless and only cosmetic.........15

3. HOW "OSMOSIS" PROCEEDS. .....................................16

3.1. The laminate of a launched hull absorbs moisture....16

3.2. Trapped substances always exist in FRP laminates..17

3.3. Water only enters along a penetrating fibre strand...17

3.4. Extended research reveals unknown damages.........18

3.5. The last missing link in the "osmosis" chain............19

3.5. Present "osmosis force" theories are impossible!..24

3.8. "Osmosis" spreads differently in deeper layers........39

3.9. Time between "osmosis" start and first blisters........45

3.10. Hydrolyse does not create de-laminating pressure.47

3.11. The "osmosis" process needs minor water supply.48

XVI

4. WHY SO MANY DIFFERENT OPINIONS? .....................49

4.1. Misleading articles and laboratory reports................49

4.2. More contaminants in "osmosis" laminates..............49

4.3. Glycol believed to be "the big villain"........................50

4.4. Dibutyl phthalate causes "osmosis"?........................54

4.5. Delaminating "osmosis pressure"..............................55

4.6. Hygroscopic materials cause delaminations.............56

4.7. Summary of common false believes:.........................57

5. "OSMOSIS" EFFECTS ON FRP HULLS. .........................59

5.1. Osmosis" is not cosmetic but severely harmful........59

5.2. Test of "osmosis" FRP before and after repair:........59

5.3. "Osmosis" often renders a hull not seaworthy..........63

6. HOW TO DISTINGUISH "OSMOSIS" TYPES..................65

6.1. Visual inspection of a test hole...................................65

6.2. Measurement of hull deformation at dry setting........66

6.3. Meter readings when IR or dehumidifier are used.....66

6.4. Moisture meter readings after open air drying..........67

6.5. Choose the proper range of a moisture meter...........67

6.6. All blisters are not a sign of "osmosis"!.....................69

6.7. "Osmosis" does not always form blisters!.................69

7. COMMON REPAIRS AND RECURRENCE RISKS. ........71

7.1. Three ingredients essential for "osmosis" to start...71

7.2. The old blowtorch method: effective but dangerous.71

7.3. Gel coat peeler effective at certain circumstances...72

XVII

7.4. Risk factors cause dry and shield methods to fail.....73

7.5. Watershield reverses water molecule flow.................73

7.5.5. Moisture movement reverses again.........................77

7.6. Importance of glycol removal exaggerated................78

7.7. Glycol in recurring blisters cause warranty refusal...79

7.8. "Osmosis" repair leaves voids in the FRP..................79

7.9. Pros and cons of grinding............................................80

7.10. Pros and cons of blasting..........................................80

7.11. Both grinding and blasting create a mess................82

7.12. Misleading statements by paint manufacturers.......83

7.13. Only one way for an owner to save costs.................83

8. HOW TO OBTAIN OPTIMAL REPAIRS............................85

8.1. At least one of three substances must be removed..85

8.1.2. Acid residues and styrene must be removed..........85

8.2. Controlled thermal treatment the only solution.........86

8.3. Voids in treated laminate must be restored................91

8.4. Epoxy resins develop an oily "skin" when curing.....93

8.5. Never apply the filler directly to the polyester............93

8.6. Avoid the use of tar epoxy for the watershielding....94

8.7. Resin and hardener must be exactly measured.........95

8.8. There are no "osmosis" shortcuts..............................97

9. TECHNICAL DEFINITIONS AND TERMS.......................99

XVIII

CHAPTER 2Recurring "osmosis". Appearance and repair

1. FALSE BELIEF IN FEW RECURRENCES .....................103

1.1. Confusing rumors about recurrence extent.............103

1.2. Average recurrence rate at least 50%.......................103

2. THE APPEARANCE OF RECURRING "OSMOSIS" ......105

2.1. Hulls covered with epoxy or other water shield......105

2.2. Hulls that have been relaminated..............................107

2.3. Hulls repaired without relamination..........................109

3. REPAIR OF RECURRING "OSMOSIS" ........................111

3.1. Repair of solvent or amine blush blisters...............112

3.2. Repair of recurrence after any type repair...............113

3.3. Repair of hulls with deep total de-laminations.......114

EXCEPTIONS POSSIBLE. .................................................115

IMPORTANT CONCLUSION FROM THE RESEARCH: ....115

1CHAPTER 1

A study of the chemical processes involvedand their effect on FRP hull strength

2

31. WHAT CAUSES "OSMOSIS"?

1.1. "Osmosis" is in reality not an adequate name.In the sixties a leading yachting magazine wrote an article

about a mysterious "illness" hitting FRP boats.In short it said that intruding water caused an "osmotic

pressure" that de-laminated the FRP layers.

This theory was repeated by a lot of other magazines andcreated the now universally accepted name "osmosis".

The basis for these articles, most likely, was the results ofcore sample testing provided at that time. In all samples from blistered hulls, the tests revealed amuch higher content of WSM:s (water soluble materials),glycol and free styrene than normal for a sound laminate.

Lacking practical field experience with the problem, anincredibly long-lived but false theory was introduced, that thegel coat blisters are formed when the concentrated solution ofWSM:s and glycol in the laminate absorbs water through thegel coat by osmosis.

The term osmosis was born and spread all around theworld!

In fact there is practically no osmosis involved in theprocess and it certainly does not create any delaminatingpressure.

Today, with the field knowledge that gallons of thesubstances in question can be found within a few m laminate,it is easy to understand, that the substances found are createdby the process and not the opposite.

4Incredibly enough some core sample testing laboratoriesstill report findings of substances which might cause osmosisdamages and thus keep the myth alive.

Also the myth has been used by the boat industry in order tomake the public believe in a minor problem caused by a fewcareless boat builders!

As late as 1988 a couple of well known boat builders wrotearticles in English an Scandinavian yachting magazines, tellingthat the osmosis problem itself was a myth and really did notexist!!

1.2. Osmosis reason: Hydrolyse of uncured resin!Five years of intense laboratory research and practical tests

on hundreds of samples from cut up osmosis hulls as well ason whole hulls gave us quite a new view of the reasons for andthe damages formed by the osmosis plague.

The main reason for the osmosis was found to be, thatboat hulls are too large to be post cured within an economicalrange.

Post curing at 80C after the initial curing in the mould isneeded for polyester resins to achieve a total curing of all thepolyester chains.

Instead of post curing, usually laminate dimensions areincreased to compensate for the 10-15% lower strength due tothe missing post curing.

There have been few considerations about the chemicalconsequences!

A not post cured laminate contains as much as 10% styreneenclosures with a skin of uncured polyester between themand the cured polyester.

5Further many of the cured polyester chains have no curinglinks in the ends, so called dangling chains

1.2.1. The styrene enclosures:The osmosis begins, when under certain circumstances

moisture enters the styrene enclosure.Water in combination with the styrene hydrolyses the

surrounding uncured polyester.The large polyester molecules split into a multitude of

smaller acid, alkali and glycol molecules which need morespace than the original ones.

This creates a pressure which forms blisters in the elasticgel coat.

A not commonly known effect of the osmosis process is,that the very small alkali molecules follow the inward capillarymovement of the moisture and cause a loss of bond betweenfibreglass and polyester.

The result is a loss of strength in the hull of 20-30%,and abnormal damage can be caused by the slightestgrounding.

1.2.2. The dangling chains:Dangling chains are the main reason for many "osmosis"

recurrences after common treatments.As mentioned before, acid e.g. pthtalic acid is formed during

the hydrolysis process.The acid is also absorbed into deeper laminate layers and

will stay there despite any kind of normal washing and drying.

When pthtalic acid molecules come into contact with curedpolyester chains attached with dangling chains, the totalchain breaks up into styrene and uncured polyester.

6The resulting mixture is extremely sensitive to hydrolysisand needs small amounts of moisture to restart the process.

1.2.3. Enough water always availableBeside moisture entering from inside condensation and the

bilges, even the most successful drying will leave at least 0.5%moisture in the laminate, enough for severe hydrolysis toproceed in spite of high quality watershield systems!

Especially in large hulls with heavy roving layers and/orsandwich core, such a huge amount of polyester can behydrolysed that several m of one or more laminate layers areno longer bound together.

After a short time this "de-lamination" fills with gallons of anacid/glycol/water mixture.

If left unattended, in a few years the laminate will have lotsof areas where only the gel coat over wet soft fibreglass is left!(See pictures)

Very often no telltale blisters will form as the pressure isevened out in the de-laminations.

Fresh water flushing of a 3mwide hydrolyse delamination.

Only soft black acid soaked fibresremain here after removal of water-shield and two laminate layers ofold repair material.

Acid fluid from deep laminatelayers wells out from testhole.

71.3. Uncured resin in laminate can be hydrolysed.To enable an "osmosis" process to start, molecules of

uncured polyester, styrene and water must come into directcontact. Fortunately, water cannot affect properly curedpolyester.

However, when moulding FRP, there are a lot of more orless uncontrollable circumstances, which cause excessstyrene to be trapped inside the cured material.

Such styrene enclosures are always separated from thecured surrounding by a skin of not cured polyester resin.

1.4. Why does the "osmosis" not start in all FRP hulls? Most "experts" explain that "osmosis" is merely a result of

faulty manufacturing. Regarding boats, water is supposed to enter into the hull by

diffusion and capillary effects and then by osmotic force bebrought into the enclosures.

Not accepting this for a truth, the HYAB organization spentmany years analysing "osmosis" hulls for some substance thatcould be the trigger of the process.

Finally in 1993 the answer was found.

1.5. Trigger found: fibre strand penetrating enclosure.Finally in 1993, studying an old laboratory report about

"leaching"(very similar to "osmosis") in FRP roofing elements,a plausible theory was found.

Even if not practically tested, it gave reasonable answers toall earlier pecularities.

8The report ascertained that there can be no osmotic forcebetween pure water and pure styrene but that water mightenter along a fibreglass strand piercing the styrene enclosure!

Practical tests performed by HYAB have proved this theoryto be valid.

1.5.1. Enclosures do not normally form around strands.Further the tests showed, that except when damp

fibreglass is used (in such cases it is well known that"osmosis" always occurs), it is not normal that enclosures areformed around a fibre strand.

Due to the high surface tension of the uncured resin, theballoon around the styrene enclosure is deformed by thesurrounding glass fibres but seldom pierced, especially whereno woven roving is involved.

1.5.2. Many FRP hulls have no trigger type enclosures.A great many FRP boats have no such fatal enclosures and

will consequently not develop "osmosis".The fibre strand being "the trigger" solves the other

confusing fact, that "osmosis" never affects the gel coat.

This material always has lots of small styrene enclosuresjust under the surface close to the water.

According to any other explanation the "osmosis" wouldhave started there!

1.5.3. Small size hulls less effected than the big ones:Looking at the published estimations of osmosis among

the total fleet, one can clearly see that the percentage of boatsaffected increases proportionally with the average hull size.

Since the 15% reported in 1980 the figures have nowreached an estimate of 40%

9The HYAB research, performed at 5 test centers in differentparts of the world, now presents a more exact picture of theosmosis frequency:

Hull size < 25 = 5%, 26- 40 = 15%, 41- 50 = 30%,51- 65= 60%.

As to bigger hulls, so far the HYAB research covers only 6yachts, all of which were affected.

Regardless of size, hulls containing woven roving mats ofover 900 g/m have a very high osmosis frequency.

If they also have core material reaching down under thewater line, the frequency seems to be 100%!

1.6. Construction faults influence process severity.If triggers are present, the number of such styrene/uncured

polyester enclosures and the situation determin the speed ofthe process.

If no triggers exist, no osmosis will develop regardless ofenclosure amount.

Below some of the most common enclosure sources aredescribed.

1.6.1. Enclosures emitted from the curing of the gel coat.

Most, not to say all FRP boats made in moulds without postcuring have styrene enclosures in the first two layers of FRP.

10

The enclosures are caused by styrene emitting from the gelcoat, which is not fully hardened when the first layers areapplied.

A laminate with up to 5% styrene/uncured polyester anddangling chain molecules in the outer laminates and very fewenclosures in the inner laminate can be considered as topclass.

The existence of trigger enclosures is pure chance, whichno boat builder can be blamed for!

If no woven roving exists in the laminate, an eventualhydrolysis process will mostly be limited to the outer 2-4 layersduring at least five years from when the first blisters werenoticed.

A peeling and "dry and shield" repair might stop furtherhydrolysis and blistering but will not restore the loss of bondbetween fibreglass and resin.

Even in a top class laminate this causes at least 15% lessstructural strength.

1.6.2. Styrene enclosures from disturbed curing.

Disturbed curing is another cause of styrene enclosures.Normally the curing proceeds uniformly right through thematerial by styrene crosslinkingthe polyester molecules.Excess styrene evaporates.

11

A number of factors can disturb this process. In connectionwith "osmosis" the most important ones are:

Draught over the curing surface causes uneven curing fromthe surface downwards.

This traps some of the excess styrene inside the FRP andcauses styrene enclosures all through the laminate.

Labour welfare acts in the eighties, demanded improvedventilation.

Many ventilation systems were built up with suction hosesarranged over the moulds.

Some manufacturers had years of production where a greatnumber of boats "inexplicably" developed "osmosis".

Later it became evident that the ventilation hoses wereplaced a foot too deep into the mold and thus caused fataldraught over the surface.

Too much accelerator, re-accelerated resin or too littlecatalyst force a small part of the curing polyester to "spit out"the styrene meant for the curing during the gelling stage.

The thus uncured polyester again forms a bubble aroundthe styrene and is accordingly enclosed in the cured FRP.

12

1.6.3. Enclosures caused by moist fibreglass.

Use of moist fibreglass in the production creates the worststyrene enclosures.

The fibres then becomes almost totally coated with styreneand uncured polyester.

Takes a long time to form gel coat blisters. Often these arelow and as big as hands.

Can also occur on deck and cabin roof areas.

1.6.4. Styrene enclosures are a property of the material.As mentioned before, post curing at 80C will rid the

laminate of such enclosures when the uncured polyester usespart of the styrene to cure and excess styrene evaporates.

It has been practically impossible, within economical limits,to mould FRP boats from polyester types used before 1990entirely without styrene enclosures.

It is just a property of the material like metal corroding orwood cracking.

No one can be blamed for this except those who used dampfibreglass or reaccelerated resin.

13

2.4. Woven roving creates more triggers.The use of roving in one of the 3 - 4 outer layers often

results in more trigger type styrene enclosures being formed.

2.4.1. Styrene surplus is divided by the roving meshes.The styrene emitted from the curing gel coat is divided by

the roving meshes.Instead of a few "normal" enclosures it forms a lot of smaller

enclosures between the roving and the next laid layer and / orin the roving meshes.

2.4.2. "Fish net action" causes penetration of enclosures.When the next layer of mat is laid or sprayed over the

woven roving mat, fibre ends are apt to protrude into themeshes like fish into a net.

Thus more styrene enclosures may be penetrated fromsuch fibreglass ends.

14

2.4.3. Roving provides a "highway".The multitude of long straight fibre strands in the roving

bundle are seldom totally wetted from the polyester resin.They provide capillaries for the hydrolysis process which,

compared to chopped strands, are like an empty highway for acar instead of passing through a medevial town!

2.4.4. More space for molecules delays gel coat blisters.Even if the process proceeds faster, capillaries along the

roving strands can accomodate many more molecules thanalong chopped strands.

Therefore the hydrolysis proceeds invisibly during longperiods and causes severe deep damages before gel coatblisters form.

2.4.5. The woven mat bond is severely disintegrated. Even in a sound laminate the bonding of the woven roving

to the next layer is somewhat disturbed by the styreneenclosures.

As the hydrolysis proceeds unnoticed for such a long time inthe woven mat, the mat will separate from the underlyinglaminate in certain areas.

This separation provides additional space especially for thebig salt and glycol molecules. Often there is enough space forsalt and glycol also to absorb lots of sea water.

If the hydrolyse in the roving is allowed to proceedunrepaired for a long time after gel coat blisters becomevisible, the separations of the woven mat will be noticeable asbig soft spots in the hull.

15

2.5. Some blisters are harmless and only cosmetic.Mainly the styrene enclosures emitting from the curing of

the gel coat are formed between the first two laminate layersor deeper. Under certain conditions they may form betweenthe gel coat and the first layer. Then the polyester will remainuncured only between styrene and laminate but not betweenthe styrene and the gel coat.

_________________________________________________________

In water temperature over 20C, styrene molecules willpass out through the gel coat blister into the sea. Water willenter, first by diffusion and then by true osmosis.

When the ideal proportions of styrene and water haveformed, the break down of the uncured polyester will start.

In this case the reaction is not retained by insufficient space.The gel coat will blister immediately.

Acid and alkali will react normally (just as by boiling testssee 4.3.1) and form glycol and neutral salts. As the alkalis areweak and the acids strong, the salts will cause some aciditywhen dissolved in water.

16

Without any surplus acids the "osmosis" process will stopwhen all of the uncured polyester is broken down. It will not spread within the laminate, but the blister maygrow to the size of a thumbnail or more due to water enteringby true osmosis.

2.5.1. Harmless "osmosis" blisters are less acid.Typically the blisters form within a few years and only in

some areas of the hull.The contents of glycol and salts dissolved in water are very

fluid and will eject in meter long squirts if a blister is ruptured. The acidity is low and the smell is somewhat "rotten" insteadof acrid.

2.5.2. Harmless blisters may exist together with the fatal.Often both types of blisters show on a hull. It is therefore

important to check blisters in different parts of the hull beforedeciding on treatment.

3. HOW "OSMOSIS" PROCEEDS.

3.1. The laminate of a launched hull absorbs moisture.Outside water pressure and interior capillary forces cause

FRP laminates to absorb water, mainly saturating thefibreglass but also water molecules can enter into themolecular sysem of the cured polyester resin (see 4.6.).

Not enough to sink the boat or even be noticed in thebilges, it makes the boat heavier.

Further the water hydrolyses the coating on the fibreglassand creates capillaries along the glass strands.

17

3.2. Trapped substances always exist in FRP laminates.Trapped styrene, small amounts of crystallized by-

products formed during the curing and sometimesmicroscopic glycol residues are found in all polyesterlaminates.

3.2.1. Styrene enclosures covered with uncured polyester.Styrene enclosures are always covered with a globule of

uncured polyester which can not be penetrated by the watermolecules.

3.3. Water can only enter along a penetrating fibre strand.Only if the styrene/uncured polyester globule is pierced

by a fibreglass strand, can water enter and cause achemical reaction.

3.3.1. Hydrolyse of polyester causes pressure.Water and styrene together causes an hydrolyzation of

the large compact polyester molecules into a multitude ofmuch smaller but more space demanding acid and alkalimolecules.

Thus a pressure is created that forces the new moleculesto pass out into the surrounding laminate.

3.3.2. Earlier theories expect molecule passage to the sea.Earlier theories expect that the smallest alkali molecules

move so fast out into the sea that no reaction with the biggeracid molecules occurs.

This is supposed to explain the formation of the pthalic acidsurplus.

18

3.4. Extended research reveals unknown damages.

AcidicBlister Fluid

No blistersdeeper than twolaminate layers

"Osmosis" laminate testpanel

Alkali residuesfound on fibers

The HYAB research team found, that the alkalis instead ofleaving the laminate were dissolved in the moisture whichalways moves inward along the fibre capillaries in thelaminate.

When examining this, it was also found, that fibreglassstrands in the inner laminate layers could easily be drawn outof the resin with a pair of tweezers!

3.4.1. HYAB tests show severe loss of strength.When comparing the results from all the tests, done by

ourselves or at independent testing institutes, figures werefound that clearly contradicted any earlier theory!

In all the tests of "osmosis" type 1 with only the two topFRP layers visibly affected by "osmosis" the samples proveto be appr. 30% weaker than an unaffected FRP.

Consequently one expects unaffected samples from thesame hull to show the same 30% loss of strength if the twotop layers have been peeled or grinded away. On the

19

contrary here all tests prove only 8 - 14% loss depending onthickness of the samples.

As the "osmosis" samples, of 8, 10 and 12 mm thickness,prove almost an equal 30% loss of strength, the total depth ofthe FRP laminate must be affected by the "osmosis"!

3.5. The last missing link in the "osmosis" chainThe immediate conclusion of the research team was, that

the loss of strength might depend on the loss of bond betweenfibreglass and resin caused by the alkalis!

3.5.1. Loss of strength caused by the alkalis.Extended tests revealed that the alkalis stay in the

laminate and causes the observed loss of strength.For the first time a complete description of the start of

the"osmosis" process can now be presented.

3.5.2. Alkalis loosen the polyester / fibreglass bond.Som very small alkaline molecules move quickly along

the capillaries between fibreglass and polyester.Together with water molecules the alkalis form a lye that

dissolves the bond between the fibres and the polyester,thus weakening the laminate considerably.

The bond mostly consists of polyvinylacetate (PVA). Alsoin sound laminates intruding water slowly hydrolyses someof the PVA into vinyl and acetic acid (so-called aging orfatigue).

This causes a 5 - 10% loss of impact and bendingresistance after a number of years (see TNO test 9.1. page34).

With the alkalis present the process is incredibly fast andthe loss of strength more than doubles.

20

The space in the capillaries is considerably widened fromthe process but still it is microscopic.

The amount of small uncomplicated alkali moleculesformed by the"osmosis" in the top laminate layers istherefore enough to affect the whole depth of the laminate.

3.5.3. Acetic acid causes the typical smell of "osmosis"The PVA substance constitutes a very small percentage

of the FRP and, in contradiction to earlier theories, theacetic acid formed is of no importance in the "osmosis"process.

Its small molecules however move very freely, wherewater is present, all the way into the outer laminates andlater into the gel coat blisters, where they cause the typicalacrid smell.

3.5.4. Glycol and salt form "border zone".A large amount of the alkalis reacts in a normal way with

acid molecules forming new bigger molecules of neutralsalts and glycols.

The type of salt differs continuously during the process.The glycol is mostly propylene or neopentyl glycoldepending on the type of polyester.

Sometimes also small amounts of hexylene glycol form.This is a very agressive solvent which can enlarge thecapillaries and speed up the process.

With the exception of the hexylene ones, the big glycolemolecules are not able to pass along the fibres.

Instead they fill the space left by the dissolved globuleand neighbouring enlarged capillaries, blocking the entranceof the surplus acids.

21

3.5.5. Formed salts and glycols cause research confusion.The large amounts of glycol and salts have caused most

of the confusion and rumours among "osmosis" specialists.

All earlier serious research reports regarding genuine"osmosis" affected FRP mention, that the "osmosis"samples contain much more water soluble materials,"WSM:s", than does the sound FRP.

It was concluded that the "osmosis" is caused by suchimpurites being embedded during the manufacturing!

To-day when we know how large amounts of "WSM:s"which can be created by the hydrolysis, it is easy tounderstand that they could not possibly originate from theconstruction!

3.5.6. Acid/styrene/water break down all enclosures.The pressure is somewhat released when the larger

molecules are formed.Still it is high enough to force water mixed styrene and

surplus acids (mostly pthallic origin ones and acid solutionsof salt in water) to move laterally in the affected laminatelayer.

On its way the mixture will encounter further styreneenclosures.

Previously the water needed a strand to enter and mixwith the styrene to start the process.

Now the mixture already contains all the substancesneeded to break down the uncured polyester globule fromthe outside and therefore also the enclosures withoutpenetrating strands are hydrolysed.

22

3.5.7. The acid mixture spreads sideways in the layer.Trying to move inward, the first entering acid molecules

meet alkali molecules that have penetrated the capillaries inthe deeper laminates.

They immediately react and form larger salt moleculeswhich effectively block further entrance of the acid/styreneresidues into deeper laminates.

As outward access is prevented by the gel coat, themixture can only spread sideways.

The "osmosis" process (which in fact has no relation togenuine osmosis) accelerates from enclosure to enclosureand the laminate layer becomes affected all over theunderwater hull.

3.5.8. The acid mixture breaks down cured polyester.As the acid content grows, it will slowly break the curing

links in the cured polyester. This releases additional styreneand uncured polyester which in turn can be hydrolysed.

3.5.9. Only when the total layer is affected, blisters form.Not until all capillaries and voids in the first affected

laminate layer are filled, the pressure is reached which canform the gelcoat blisters.

3.5.10. Size of blisters is no indication of damage state.Size or frequency of blisters does not indicate the severity of

the breakdown process, only the properties of the gel coat.

Thick or solid gel coat often shows more and bigger blistersthan thin and porous.

It is just a balance between pressure from the inside, timeneeded for the residues to pass through the gel coat and gelcoat elasticity.

23

3.5.11. Blisters may disappear after a while.At this stage the "osmosis" process slows down

considerably, because only cured polyester remains to "feed" itin the affected layer.

More residues leak through the gel coat than what arecreated in the process and the pressure decreases.

In many cases the blisters disappear until also the secondlayer has been totally affected.

3.5.12. Weight of hull forces water into the neutral zone.Until now the pressure from the process has been higher

than the pressure from the hull on the outside water.As the pressure decreases, water as moisture will pass

through the gel coat into the "water hungry" salt- and alkalisolutions in the inner capillaries.

The glycol molecules on the contrary are so big that theycan not migrate into the capillaries. They are "imprisioned" inthe original styrene enclosure space.

Just a few added water drops will be enough to fill anysurplus space and stop further water entering into the glycolfilled enclosures.

The pressure involved can not cause any delaminations!An exception is where a layer by construction has been

allowed to cure before the following layer was applied.Then often the bond is so weak, that the total outer layer

can be de-laminated from the inner laminate also from thismoderate pressure.

24

3.5. Present "osmosis force" theories are impossible!

3.4.1. Explanation of osmosis.

If solutions of different concentration are separated by apermeable wall, solide with holes or an absorbing substrate,they will pass into each other until the concentration equals.

No pressure or different amounts on the two sides will arise.

If they instead are divided by a material which will onlypermit molecules of the solvent, in this case water, to passthrough (semi permeable membrane) only pure watermolecules from the low concentration side will pass into thehigh concentration solution until the concentrations equals.This phenomena is called osmosis.

If a certain pressure is applied to the high concentration sidethe passage will stop. This pressure is called the osmoticpressure and depends of the difference in concentrationbetween the two solutions.

Big difference = higher pressure, low difference = lowerpressure.

25

In the first example above the contrapped air will becompressed by the water molecules entering the highconcentrated solution until the osmotic pressure is reached,then the process stops.

The second and third examples shows how the weight ofthe high concentration solution creates the pressure wichstops the process if the two surfaces are not closed in.

Higher concentration = higher level before stopping.

3.6.2. Gel coat is not a semi permeable membrane.The osmosis is never relevant in a fibreglass hull. The gel

coat is an absorbing substrate (even if it is less absorbing thanthe laminate) and will let water out or in as moisture togetherwith most of what is dissolved in it.

As well as lower concentrations from the outside will passinto the laminate, eventually higher concentrations inside thelaminate will pass out until the concentration equals.

If the water on the outside, like under the weight of a boathull, is under pressure, moisture will be pressed through thegel coat into the laminate.

After that the laminate has been saturated and eventuallydelaminations filled, the same amount of moisture which

26

enters through the gel coat will evaporate from the inside ofthe laminate.

This will never cause any pressure within the laminate orbetween laminate and gel coat as supposed by earlierpresented theories.

3.7.Chemical reasons for polyester to hydrolyse.Lower esters can be formed just by mixing acid and alkali

molecules under de-hydrolyse e.g. removal of a vater moleculeduring the process.

For high molecular esters used for polyester resin differentglycols are used instead of pure alkali.

Also those glycols are de-hydrolysed during themanufacturing process.

During the mixing with the acids used further de-hydrolysewill occur.

Below some shematic descriptions of the two most commonesters used for boat construction, the ortho and the iso esters.

3.7.1 The ortho ester molecule.The ortho ester is formed by mixing one phthalic acid

molecule with two propylene or ethylene glycol moleculesduring de-hydrolyse of two water molecules.

The orto ester now has one molecule group on each sidewhere only addition of a water molecule will form glycol again.

27

Also the center group can be hydrolysed by one or twowatermolecules. The result will be a mixture of phthalic acid,alkalis, salts and lower and higher glycols.

3.7.2. The iso ester molecule.By mixing two of the smaller isophthalic acid molecules with

only one bigger neopentyl glycol molecule the iso ester formswithout dehydrolyse and should be less apt to hydrolyse.

Only one side of the molecule has a group which can behydrolysed to glycol.

However if we look closer at the iso- phthalic acid moleculeit is just a de-hydrolysed normal phthalic molecule!

28

Therefore the other side can be hydrolysed to acid and thecenter can be hydrolysed just as the ortho ester.

The iso polyester is better than ortho in many aspects. Less uncured enclosures with lower amount of "triggers"might slow down the start of the hydrolyse but when it starts,double the amount of the agressive acids forms. The laminate often gets worse so called "osmosis"damages than the ortho types.

Earlier iso types also formed more of the dangerous"dangling chains", as any end of the molecule could add to theend of the next molecule when the polyester chains formed.

Later types have an oxygene or a carbon dioxide moleculeadded to the iso phthalic acid e.g. terepthtalic acid whichreduced this problem.

29

3.7.3. Polyester production.As the name indicates, the polyester molecule is formed by

a large number of ester molecules "glued" together as a chain.For most boat construction polyesters the "glue" will be maleicanhydride molecules.

30

As longer the chains as more solide the polyester will be atroom temperature. The boat construction polyester moleculesmust have a certain "high molecular" length in order to obtainsufficient hardness after the curing.

The "virgin" resin has a far too high viscosity to be used forlaminating and a solvent must be added.

The styrene which must be added in a certain amount forthe curing process is also a suitable solvent.

By adding a surplus amount of styrene, the viscosity isbrought down to a suitable level.

All of the substances used when creating the polyester resinare highly reactive and instable. So is also the resin itself.

Especially are all of them "water hungry" due to the manydehydrolysing steps involved during production.

Normally the curing and post curing will make the materialstable and water resistent.

Now the post curing is abandoned and as mentionedbefore, lots of uncured polyester, styrene and other reactionproducts will remain in the laminate.

3.7.4. The polyester curing process.When a small amount of methyl ethyl ketone peroxide is

added to the polyester resin, a reaction starts, where pairs ofmolecul chains are cross linked by MEK and styrenemolecules. The molecular weight raises andthe resin becomes hard e.g. cures.

Only part of the styrene is used for the cross linking. Therest is supposed to escape from the laminate during the curingperiod.

31

After the normal temperature curing in the mould, the resultis demonstrated in the above schematic illustration.

Only part of the surplus styrene escapes. The rest formssmall blisters surrounded by a thin skin of uncured polyesterentrapped mainly in the outer laminate layers.

Some polyester chains remain incomplete cross linked.Most of them lose the normal glycol end which is replaced witha phthalic acid molecule e.g. "dangling chains".

32

The free glycol, not used styrene and MEK will remainwithin or close to the faulty molecule for about a week.

If within this week the laminate is heated for a short time to80C, the skin around the styrene enclosures as well as theincomplete cross linked chains will cure properly and allexcess styrene will evaporate.

If the laminate is not post cured, uncured polyester andproducts formed by the other remnants will invite water to startthe "osmosis" e.g. hydrolyse.

The acid formed by the hydrolyse will break the danglingchains into substances, which in turn can be hydrolysed. Whenacid and styrene are sufficiently concentrated, also properlycured molecules will be broken and hydrolysed. The normallyhard laminate turns into soft wet fibreglass.

33

3.7.5 Illustrations of first stage FRP hydrolyseShown below a schematic cross section of the gel coat and

the two first laminate layers from an osmosis affected FRPlaminate.

Note! The capillaries along the glass fibres are extremly enlarged in order to provide visibility!

Fig1. Inside the gel coat the laminate consists ofcured polyester reinforced with fibreglass mats .

Variable numbers of styrene enclosures covered by alayer of uncured polyester will always remain after thecuring. Normally they cannot cause any harm.

The displacement pressure slowly forces sea water throughthe gel coat. The molecules collect to moisture in thecapillaries along the glass fibres.

High water content in the outer layers, low content in those further in. The capillary force transports the

moisture inward right through the laminate. The same amountthat enters from the outside evaporates from the inside. Thepercentage of moisture in the laminate stabilises at a fixedlevel depending on laminate quality (less than 1% up to over!0%) without any hydrolyse problems.

34

Fig 2. If the laminate contains one or more styreneenclosures penetrated by fibreglass strands, the scene differs:The water, which normally cannot penetrate or affect theuncured skin around the enclosure, enters into the enclosurealong the penetrating strand.

Here the inside of the skin is weakened by the styreneand can be attacked by the water. The uncured polyester ishydrolysed into smaller but more space demanding acid andalkali molecules.

Fig 3. In need of more space a part of the small alkalimolecules are forced out into the capillary system before theycan react with the acid. They mix with the inward moving water

35

to form a lye and spread very fast along all capillariesinto the innermost laminate. The lye affects the bindingbetween glass and polyester and the laminate getssubstantially weaker.

The remaining alkalis react with acid and form salts andglycol within the enclosure cavity and also clog theinward capillary openings.

A surplus of acid and styrene remains and is forced out intothe capillary system in the outer laminate, where they mix withexisting water. The acid mixture only moves sideways.Inward movement is stopped by salts and glycol. The outsidewater pressure prevents escape through the gel coat.

Fig 4. When the acid mixture reaches the next enclosure, itcontains styrene enough to weaken the outside of the skin.

Therefore, once the process has started, all enclosures inthe affected layer are hydrolysed.

This highly accelerates the process.

Further the phthalic acid part of the mixture is able to breakup the curing links in the cured polyester. This is a very slowprocedure but it will produce new amounts of styrene anduncured polyester to feed the hydrolyse. This property of theacid is also the main reason for the re-currency of theosmosis after a repair!

36

Fig 5. The acid mixture spreads along the fibreglass strandsfrom enclosure to enclosure in the outer one or two laminatelayers all over the underwater hull.

Once started, this osmosis process phase proceeds veryfast. In a few months the affected layers can be totallydegraded without any visible sign!

Already now the alkali lye, mentioned in fig 1, has spreadwithin all the inside layers.

The binding between fibreglass and polyester is severelydamaged and the hull has lost 20-30% of its strength.

37

Fig 6. With no more expansion space along the fibres andwith all inward directed capillaries blocked by salts and glycol,the process pressure increases.

It will totally stop water from entering through the gel coatand instead cause the acid fluid to seep out. As this is notenough to even out the pressure, the elastic gel coat will nowform the characteristic blisters.

At this stage the blisters will seldom be bigger than 1 - 15mm. The size actually depends on the gel coat quality, not onthe state of the hydrolyse.

A bad gel coat allows more pressure to escape and formssmaller blisters than a good gel coat.

If bigger than 15 mm blisters are found, the process mostlikely has reached into deeper laminate layers.

Key to colour coding

38

Fig 7. When all uncured polyester in the outer layer(s) hasbeen hydrolysed, the process slows down and no morepressure forms.

Very often the gel coat is elastic enough to let the blistersdisappear until deeper layers are affected.

During this phase of the process pure outside water willpenetrate the gel coat and dilute the salt solution on the inside.

The original inward movement of fluid in the capillaries isrestored and also the fatal acid/styrene mixture follows thewater into the next laminate layer.

Thus the hydrolyse process will repeat layer by layer untilat last only wet soft fibreglass remains instead of the solid hull!

Note! Especially if moist fibreglass or heavy woven rovinghave been used in the construction, the hydrolyse can alsostart .in an enclosure deep inside the laminate!

Then very severe damage may occur to the laminate beforeany blisters show.

39

Large areas of an inside laminate layer (sizes of 1 m arenot uncommon) can be totally hydrolysed

.The fibreglass mat remains alone without attachment to thesurrounding layers (total de-lamination).

3.8. "Osmosis" spreads differently in deeper layers.

3.8.1. Typical mushroom shaped damages form.When the "osmosis" expands into deeper laminate layers

the process alters considerably.Instead of spreading equally inside a total layer it forms the

typical mushroom shaped spot damages well known by"osmosis" repair operators.

3.8.2. More available space reduces the acid.The hydrolyse process as before creates large amounts of

acids, alkalis, glycol and salts.As much of the inside capillary space is already occupied by

alkalis from the top layer process, less of the newly formedalkalis can escape this way.

Instead they react normally with the acids. This results inless acid surplus and more salts.

Further, when the mass of new small molecules need morespace, they can freely move out into the already degradedlaminates.

3.8.3. Shrunken gel coat blisters become visible again.The hardly noticeable rise of pressure created from the new

molecules will be neutralised by the elastic gel coat blisters.Earlier "disappeared" blisters will be visible again and

remaining blisters will grow a little.

40

3.8.4. Mushroom stem is formed.The acid and/or styrene can only penetrate the neutral zone

along one or more fibre strands which also penetrate thepolyester tie-coat and into the next fibreglass layer. Thepassage will be widened as cured polyester slowly ishydrolysed.

Since the process pressure is more or less equalised theaggressive fluids are not forced to move along the lateral fibrecapillaries like before.

If no styrene enclosures are situated in the veryneighbourhood, the "osmosis" will go on hydrolysing curedpolyester deeper into the laminate and also slowly widen the"stem" as shown in the lower part of the illustration. (Fig. 8)

3.8.5. Mushroom caps are formed.If styrene enclosures are present near the "stem" in a

fibreglass layer they will be hydrolysed like in the start in theouter layers and the process will spread laterally again.

However much of the pressure will leak outwards and alsoless acid surplus will form. The spread of the damageproceeds much slower than in the outer layers.

Instead of the whole laminate layer being destroyed, onlysomewhat circular damages form like mushroom caps as seenin the upper parts of the illustration (Fig 8).

3.8.6. New neutral zones make mushrooms fork.New neutral zones form on top of the "caps" which cause

new "stems" to enter the underlying layers and so on.Often two or more "stems" rise from each "cap" as also is

illustrated (Fig 8).

41

3.8.7. Abundant enclosures cause wide "de-laminations".In laminate layers with large amounts of styrene enclosures

all of the polyester within a "hat" area can be degraded andmake it look unwetted or de-laminated.

3.8.8. In due time the hull will render completely soft.Without exception sooner or later an "osmosis" affected hull

will be rendered to soft to be used.When gone so far it is also too late for a repair!

3.8.9. Illustrations of second stage FRP hydrolyseTheosmosis, or let us from now on use the more proper

word hydrolyse, shows a different chemical behavior when itproceeds into deeper laminate layers.

Also the laminate lay up type causes significant differences.

Fig. 8 Visualises three different types of inwardmovement of polyester hydrolyse.

In reality they are of course considerably more spacedapart from each other.

42

Second stage damages in this type of laminate might besuccessfully repaired, like first stage ones, using a commondry and shield method, if all of the mushrooms are found,ground out and re-laminated.

As the hydrolyse will spread again, if just one tiny capremains somewhere in the laminate, the risk for recurrence isconsiderably high.

The uncured polyester, acid, glycol and dry salts in the caparea will not cause any turn of the moisture meter scale, butthe hydrolyse will start as soon as some moisture enters again.

With the outside covered by an epoxy or other type ofwater-shield, the cold wall principle will cause insidecondensation to enter into the laminate.

Fig 9. Visualises how the inside of woven roving matlayers are totally hydrolysed and de-laminated.

43

When a mushroom stem reaches into a woven roving layer,acid, styrene and moisture will spread very fast into largeareas of this layer. This spread is not caused by pressure but is due to capillaryforces along the mass of extensive fiber strands.

Instead of forming a local mushroom cap damage, in acouple of years one single stem can cause the whole layer tobe hydrolysed and totally de-laminated from the next layer!

In laminates with multiple roving layers, such de-laminationsoccur in many of the layers.

The de-laminations are often square meter large andoverlap each other.

Especially in hulls earlier repaired for osmosis, the de-laminated roving layers will totally even out the processpressure and no tale telling blisters are formed in the coating.

The boat owner does not become suspicious until the boatbehaves peculiar at sea or visibly deforms by drysetting. Bythen probably the repair warranty has expired.

Already in a 50 hull it is quite common that 20-30 liters ofaggressive fluid are released if a test hole is drilled in an areawhere the moisture meter shows just a slightly higher readingthan normal.

If the damage lays deeper than 6-8 mm from the surface, aninstrument with close spaced electrodes like the Sovereign,used by many surveyors, does not indicate any moisture at all!

Analysiss of test cores give very different results dependingon laminate quality and type of polyester used.

The worst fluid containing de-laminations are found in bigortho polyester hulls made in hot climate areas and where

44

some laminate layers had time to cure totally, before the nextlayer was applied.

Such layers de-laminate so early during the hydrolyse, thatinstead of a phthalic acid surplus an alkaline maleic anhydridesurplus forms.

The also formed dimetyl phthalate salt, will be diluted byentering water and then become slightly acid.

In total the fluid will be anything from slightly acid to slightlyalkaline, where alkaline signifies a bad lay up and acid a betterone.

This might seem contradictory, but the inward spread of thehydrolyse proceeds slower in a bad laminate than in a betterone, even if the resulting de-laminations become worse.

It might take 10 - 15 years before the inner half of thelaminate is affected!

Alkaline readings of fluids from inside de-laminations orpockets must not be confused with alkaline readings of fluidfrom blisters in an epoxy watershield coating! Such blistersoriginate from curing problems in the epoxy coating layers andcontain amine carbonate. They have no connection with thepolyester hydrolyse.

Regarding iso polyester hulls we have still only foundhydrolyse in roving type laminates.

Here the laminate layers are better bound to each other.Much more surplus acid forms, and all of the roving layers rightthrough the laminate are totally affected within a few yearsfrom the start.

The bond between the layers is severely weakened but totalde-laminations deeper than 4 -6 mm from the surface are rare.

A drilled test hole mostly only gets moistened by a fluid witha high content of phthalic acid.

45

However stresses on keel and rudder often cause theweakened areas around them to de-laminate completely.

Any attempt to repair a second stage hydrolyse damage in aroving type hull, using current methods, is deemed to fail.

Dry and shield repair can even speed up the process byreversing the capillary movement of moisture in the laminate.

Use of vacuum pumps to empty the de-laminations and/orpeeling and applying of 4 - 6 mm of new laminate might resultin a dry reading on a Sovereign or similar moisture meter andan approval from a surveyor, but the process will proceedinside the repair.

Usually it can soon be recognised by an acid smell in theinterior of the hull.

3.9. Time between "osmosis" start and first blisters. The speed of the "osmosis" process depends on material

quality and surroundings.

It may take anywhere from a few months to many yearsfrom when the "osmosis" starts in the first styrene enclosureuntil the first blisters appear.

3.9.1. A total layer may be affected within a few months.A bad FRP with many styrene enclosures may have the first

layer affected totally around the underwater hull within a fewmonths from the start, while a good FRP quality can need 10years or more.

46

3.9.2. Temperature and type of water important factors.The environment has a big impact on the "osmosis". The

hydrolyse process is considerably slower at temperaturesbelow 15C and practically stops below 10C.

Fresh water, especially when somewhat acid, causes afaster break down, than salt sea water.

3.9.3. Content of butyl alcohol is another important factor.The curing process always leaves residues in the laminate.

One of them, dibutyl phthalate, reacts with water and formsbutyl alcohol which is highly hygroscopic.

The butyl alcohol partly moves into the gel coat blisters,where it rapidly vanishes out in the water and partly into theneutral salt zone.

The more alcohol in the zone, the faster the acid mixturecan penetrate it.

3.9.4. Patent pending test technique developed.The HYAB team has developed a test technique, which

shows exactly which layers are affected and to what extent.The operator can immediately tell the owner if the hull can

be repaired with lasting result, what it will cost and exactly howlong time the repair will take.

3.9.5. Patented repair technique developed.The research made it also possible to create a reliable,

patented repair technique, the HYAB Osmocure which isdescribed in the book

"Osmosis" How to repair hydrolyse damages using theHYAB Osmocure technique" by the same author.

47

3.10. Hydrolyse does not create de-laminating pressure.There exists a very strong belief among "osmosis"

specialists, that an "osmosis pressure" causes the FRP to de-laminate.

Osmotic pressure is instead the pressure needed inside adiluted substance to prevent pure solvent molecules fromentering through a dividing membrane.

Even if osmosis should have been involved the pressurewould not be high enough to cause delaminations.

The substances formed by the hydrolyse for certain do notcreate such high pressure.

3.10.1. De-lamination involved is a chemical degradation.By grinding, IR heating or use of the HYAB lance, often big

areas of fibreglass without polyester, looking as if they had notbeen sufficiently saturated in production, are found.

In reality they are caused by the hydrolyse process.

3.10.2. Heat is the best detector of the damages.When using heat type equipment, many operators do not

understand, that they detect already existing damages.They believe that they have caused the "de-laminations"

with the heat and therefore refrain from the use of the onlyequipment that provides a solution to the "osmosis" problem!

3.10.3. Forced or "osmosis" de-lamination?If an actual de-lamination caused by force is inspected, one

can clearly see lots of ruptured fibreglass strands.By the "osmosis de-laminations" no such ruptures can be

found. One must understand, that the dry fibreglass instead iswhat remains after that the "osmosis" process has destroyedall of the polyester!

48

"Osmosis" de-laminations

Smooth surfaces

Forced de-laminations

Surface fibers torn apart

It is absolutely necessary to open up and restore all suchdry or fluid filled de-laminations during an "osmosis" repair.Some can be square meters large!

3.11. The "osmosis" process needs minor water supply.When "osmosis" develops in the inner laminates it can

proceed with just a minor water supply! Once started in an enclosure it draws moisture from a

waste part of the surrounding FRP.

Experience from re-doing a substantial number of boats,previously treated for "osmosis", has proved, thatcondensation on the inside of the hull is sufficient to maintainthe process.

Especially in yachts used year-round as quarters for manypersons, "osmosis" often recurs within a year.

Note that many types of epoxy watershields do not allow therestarted "osmosis" to form blisters!

49

4. WHY SO MANY DIFFERENT OPINIONS?

4.1. Misleading articles and laboratory reports.Besides misleading articles mentioned there are lots of

other reasons for the confusion.Before the HYAB research there existed very few

laboratory test reports of actual "osmosis" affected hulls otherthan of small brick drilled samples

.Only a few reports about the contents in the gel coat

blisters can be found, some being analyses while others aretheoretical. Formulas several lines long, often different andcontradictory are presented.

This is can be understood by a chemist but is veryconfusing for people involved with common boat production orboatyard labour.

In fact not only does the blister content differ in each hull,but they continuously change because of the instability of thechemicals involved.

4.2. More contaminants in "osmosis" laminates.Many analyses of actual "osmosis" laminates reveal a

much higher content of WSM:s, glycol and free styrene thannormal for sound laminates.

Instead of understanding that most of the contaminants arecreated by the hydrolyse, the laboratories involved state, that"osmosis" is a consequence of too much contamination of theoriginal laminate!

50

4.3. Glycol believed to be "the big villain of the problem".Minimal amounts of glycol residues can sometimes be

found in sound FRP laminates.

One of the oldest theories about "osmosis" claims, that theprocess starts when water is attracted by this glycol.

This is supposed to explain why only some FRP hulls areaffected.

To prove the theory, laboratory reports from resinmanufacturers are presented.

4.3.1. The tests refer to boiling of unaffected laminate. Such tests are all of the type, where samples of FRP

unaffected by "osmosis" are boiled in distilled water, which isthe fastest way to make water effect a polyester glasslaminate.

The boiling causes the styrene enclosures to expand and torupture the covering "balloons" of not hardened polyester.

4.3.2. Hot polyester is plastic and permeable. As with all plastic materials the polyester softens at a

certain temperature.For polyester used in boat manufacturing this temperature is

about 80C.

At boiling temperature cured polyester is soft enough toprovide practically free passage and space for the moleculesformed, when the mixture of water and styrene breaks downthe uncured polyester.

51

4.3.3. Alkalis can not escape like they do in an actual hull. At boiling temperature all the acid molecules react with the

alkali ones.The reaction forms only neutral products, mainly propylene

glycol, glycol esters and salts.

Due to the heat this process is very fast and completelydifferent from the "osmosis" reaction in a FRP hull.

4.3.4. No aggressive acids remain after a boiling test.As the acids that create the esters are strong and the alkalis

mild, the salts formed will turn slightly acid when dissolved inwater.

There will be no surplus phthalic acid like in an "osmosiscase, only small amounts of acetic acid from the binding agent,which is also dissolved by the boiling and the entering water.

Even If the boiling proceeds for hours, the properly curedpolyester is not seriously harmed.

Also all reports mention that the breakdown stops when theboiling is finished.

4.3.5. Blisters formed by boiling have a different content .The molecules formed by the boiling also need more space

and are forced out of the FRP.They cause blisters in the gel coat similar to the "osmosis"

ones but with totally different contents.

The hot gel coat is very plastic. Most of the surplus, like thestyrene, is gaseous and escapes easily.

Only fluids, like glycol, diluted salts and a small amount ofacetic acid, remain in the blisters.

52

4.3.6. No "osmosis" starts if samples are left in cold water. If the boiled test samples are left in cold water for long

periods, nothing happens except that the blister content isdiluted by water.

No de-laminating forces or typical "osmosis" damagesoccur.

This confirms the HYAB conclusion, that no "osmosis" canstart without the presence of uncured polyester, styrene and /or free phthalic acid.

4.3.7. Styrene free laminate is "osmosis" resistant. The water resistance of cured polyester became very

evident at the HYAB 7-hour boiling test of repaired "osmosis"damaged laminates (see 5.2).

Samples repaired in such a way, that they still containedstyrene and acid residues were severely affected.

Samples where all styrene and acid had been removedwere not affected at all.

4.3.8. Boiling tests have no relevance to "osmosis". Checking the difference in weight before and after boiling is

a very adequate test for determining the water resistantqualities of different polyester laminates, but it has norelevance whatsoever to the "osmosis" process.

4.3.9. Boiling tests are easy to misinterpret.The boiling test reaction proceeds exactly as expected

theoretically and the boiling procedure also creates blisterscontaining glycol.

It is natural for a reader having "osmosis" in mind tomisinterpret the reports, especially as some of them alsoprovide theories (but no tests) why boat hull blisters insteadcontain acids:

53

Theory 1: Some of the alkalis vanish into the water beforereacting with the acids, causing an acid surplus.

Also the HYAB team believed in this theory before ourlaboratory tests proved it to be wrong.

Theory 2: The acidic reaction is caused by acetic acid whichis confirmed by the typical smell.

This must evidently be wrong as we have found the worstand most acid "osmosis" among old hulls where methylacetate bound damp mats have been used.

Further all analyses of the blister fluid, despite the differentformulas, show phthalic acid to be the basic ingredient.

It is true, that the strong acrid smell found in the blisters iscaused by small amounts of acetic acid from the PVA binder.

4.3.10. Descriptions of resin production add to confusion. Another cause for believing in the "glycol" theory is, that

most literature about polyester manufacturing does notmention anything about alkalis and the de-hydrolyse of watermolecules.

Only the mixing of semi-manufactured propylene glycol withphthalic acid and maleic anhydride is mentioned.

It is then easy to believe, that if too much glycol is added,this may result in excess glycol enclosures in the laminate.

54

4.3.11. The temperature difference hull contra sample. It seems strange, that no report mentions the different

conditions between a boat hull in cold water and the testsamples in boiling or hot water.

One explanation may be that most of the tests are relatedto the manufacturing of architectural FRP panels, where thesun would provide the heat needed to keep the process goingand rain and humid air the moisture needed.

4.3.12. Misinterpretations of boil-test reports."Osmosis experts" attempting to explain "osmosis" often

base their theories on boiling test reports.

In spite of being both confusing and contradicting, theirtheories are forwarded to and believed by yards and boatowners.

Books written only a few years ago by established"osmosis" surveyors claim that "osmosis" starts in glycolenclosures in the FRP and are often used as a reference.

4.4. Dibutyl phthalate in the laminate causes "osmosis"?Some reports mention that small amounts of dibutyl

phthalate are formed as the resin cures.Trapped in the laminate it is dissolved by intruding water

into phthalic acid and butyl alcohol, ingredients commonlyfound in gel coat blisters.

Dibutyl phthalate is therefore claimed to be the cause of"osmosis".

In reality the amount of dibutyl phthalate in a hull would notcreate phthalic acid enough to fill more than a couple of thehundreds of blisters found on a hull.

55

4.5. Delaminating "osmosis pressure".In spite of being the most incredible, the osmosis pressure

theory is the most widely spread among yards, boat ownersand journalists.

4.5.1. Long squirts of osmosis fluid prove high pressure. Some publications mention that "osmosis pressures of

20.000 psi and more" are created in the laminate and causede-laminations.

This is claimed to be proved by the long squirts of fluidemitted from punctured blisters.

The actual pressures are actually more in the range of 2 psi.The pressure in a toy rubber balloon filled with water is similarto the pressure involved.

If the balloon is punctured with a needle, the water will squirtquite a few meters!

Probably the term molecular pressure has beenmisunderstood and interpreted as osmotic pressure (See4.6)

4.5.2. "Osmosis pressure" is misunderstood. Osmosis, or more correct osmotic pressure, is thepressure needed on the most concentrated side to preventmolecules of pure solvent from penetrating a semi permeablemembrane. separating two solutions of varying concentrations.

Note the word semi permeable which means, that onlymolecules of the solvent can pass through it.

4.5.3. Gel coat is an absorbing substrate, not a membrane. As earlier mentioned the gel coat is an absorbing substrate

which can absorb moisture.

56

Sea water together with most of the substancesdissolved in it can pass through. In or out depending on thesurroundings

4.5.4. No true osmosis is involved in the blistering. Instead the difference in concentration of soluble

materials in the moisture inside the laminate and the sea waterwill even out by sea water entering through the gelcoat and thesame amount of the higer concentrated solution migrating outinto the sea.

This will neither cause any pressure nor any blistering!

4.6. Hygroscopic materials cause delaminations.Almost all popular "osmosis" literature claims that

hygroscopic materials, preferably glycols, attract water andcreate high pressures.

Such high pressures are only involved on the molecularlevel when water molecules enter the molecular spacebetween the molecules of the cured polyester.

If the molecular space is totally filled with water molecules,the molecular gravity in the polyester is disturbed and thematerial contracts with high force (up to 20.000 psi) andspeed.

The water molecules are too entrapped to escape, and thepressure brings them together into small water droplets whichin turn cause circular cracks in the contracting polyester wherethey then remain ( so-called "disk cracking").

This has nothing to do with the start of "osmosis" andfurther the cured polyester in a hull seldom becomes saturatedto such an extent.

57

4.6.1. Hygroscopics only absorb what is space for.Squeeze a sponge in your hand and submerge it in water.

Did it force your fingers to open???Or presume that your car engine is filled with a highly

concentrated glycol solution and that the expansion tank isrefilled with pure water.

Does the engine break in pieces?

4.7. Summary of common false believes:

4.7.1. Glycols cannot cause or speed up "osmosis".The insignificant glycol impurities that exist in not post cured

FRP laminates, play no part in the "osmosis" process.Just like the trapped styrene they are covered with a balloon

of uncured polyester which water molecules are unable topenetrate.

Even if water for any reason were to mix with the glycol,neither "osmosis" or any kind of pressure could be the result.

4.7.2. So called W.S.M:s cannot start "osmosis".Small oobalic napthenate and dibutyl phthalate impurities

may be dissolved by the water but cannot start the "osmosis"process.

The butyl alcohol can speed up the inward penetration ofaggressive "osmosis" residues.

4.7.3. Dissolved PVA binder only provides the smell.The PVA binder is dissolved by intruding water, but the

acetic acid produced has nothing to do with the "osmosis"other than providing the smell.

58

4.7.4. Osmotic or hygroscopic forces cause no de-laminations.

At no point of the "osmosis" process are pressures formedhigh enough to cause de-laminations. Rumours about the highpressures originate from reports about "disc cracking".

4.7.5. Sea water in de-laminations kills osmosis theory.During the HYAB OsmoCure test period 1993 - 1997,

hundreds of fluid filled de-laminations were found and repairedon hulls earlier treated by common dry and shield methods.

The fluid in the de-laminations has in every case consistedof acid and glycol diluted by large quantities of salt sea water.

With osmotic forces involved, only fresh water should haveentered.

This proves both that there is no osmosis and that theepoxy water-shield is not sufficient to stop the water frombeing absorbed by the laminate!

59

5. "OSMOSIS" EFFECTS ON FRP HULLS.

5.1. Osmosis" is not only cosmetic but severely harmful.Many boat owners believe that "osmosis" is only a cosmetic

problem with superficial blisters.HYAB and other research tests prove that it is severely

harmful.

5.1.1. 30% loss of strength before gel coat blisters form.The truth is, that already an "osmosis" type 1 causes about

30% loss of strength in the FRP due to the alkali penetration.

This initial loss of strength will not be recovered by a repairwithout a heat treatment, (e.g. HYAB or IR heating) asindicated by the following test results.

5.2. Test of "osmosis" FRP panels before and after repair:Blow energy (J) needed for a 6 mm radius point to penetrate

a test sample 80 mm in diameter and 10 mm thick wasdetermined.

The black bars in the diagrams below represent top andbottom results from 4 samples of each kind tested. Theaverage value of each test is cross marked on thecorresponding bar.

Only FRP originating from PVA bounded fibreglass mat hasbeen used.

FRP with methyl acetate bounded fibreglass is generallyclaimed to lose less strength.

As most FRP hulls are made from the PVA bound mat,there was no other object available to prove this theory.

60

Samples of different types of "osmosis" to be tested havebeen of various thickness and age.

The values obtained have been adjusted forcorrespondence to the average 10 mm test thickness and thecharts are estimated to be > 85% accurate.

"Osmosis" of type 2 and 3 are impossible to cover with afew test samples.

The unfilled bars beside the filled ones show an estimatedrange based on field experience.

A FRP hull that can be penetrated by a force correspondingto a blow energy of J 27 in this test must be considered as notseaworthy.

To achieve an approximation of "late stage osmosis", thetest samples are boiled 7 hours in distilled water.

This corresponds to about 5 years use in 15 C normal saltseawater.

61

TEST NR 1: Different types of "osmosis"

TEST NR 2: "osmosis" type 1 after repair

62

TEST NR 3: "osmosis" type 2 after repair

TEST NR 4: "osmosis" type 3 after repair

63

5.3. "Osmosis" often renders a hull not seaworthy.Analysing the above tables indicates that most types of

"osmosis" will within a few years time bring the strength of ahull to below seaworthiness.

Especially fin keeled sailing yachts will be considerablymore vulnerable to grounding.

64

65

6. HOW TO DISTINGUISH "OSMOSIS" TYPES.

6.1. Visual inspection of a test holeBased on the new understanding of "osmosis", its varying

forms can be easily identified prior to the repair.

Use the tip only of a drill bit with the same diameter as thelaminate thickness to drill a conical hole through 2/3 of thelaminate. If the laminate is cored, drill just into the core.

Drill about a dozen such holes around the underwater hullwhere the moisture readings are higher than average.

Wait for about an hour or heat around the holes with a hotair gun.

The conical walls in the hole will then show darkdiscolouring of any affected laminate layer.

Mostly in large hulls fluid may seep or well out (useprotective eye glasses when drilling).

Place a strip of litmus paper against the inside of the hole. Ifthe pH is below 4.5 and no fluid seeps out, the discolouredlayers are affected but not yet de-laminated.

If fluid of this acidity seeps out, the layer has lost somebond.